Calculus Accumulation-Based Inspection Reading Method for Substation Pointer Meter

Outdoor pointer-type instrument working conditions are complex, the dial is susceptible to environmental pollution, and the readings are difficult to identify accurately online. A two-dimensional convolution and calculus accumulation-based instrument reading recognition scheme is proposed, and a weighted multi-feature matching algorithm is adopted to accurately obtain the center coordinates of the rotating axis, to improve the success rate of the instrument image recognition. The target pointer position is obtained by polar coordinate transformation and calculus accumulation of the dial image, and automatic readings are realized. Through the experimental comparative analysis of outdoor dust blocking, liquid droplet blocking, leaf blocking, and different lighting conditions, the results show that the algorithm has good robustness, almost independent of the external environment, and the recognition success rate reaches more than 98%. It provides a good reference for the unattended pointer instrument.